More data on 7.2-RELEASE "hangs"
John Baldwin
jhb at freebsd.org
Wed May 13 16:52:33 UTC 2009
On Wednesday 13 May 2009 12:34:39 pm Marc G. Fournier wrote:
> On Wed, 13 May 2009, John Baldwin wrote:
>
> > On Wednesday 13 May 2009 3:09:33 am Marc G. Fournier wrote:
> >>
> >> Don't know if this helps with anything, but it just hung after 2days
again
> >> ... nothing on the console ... top process running at the time shows the
> >> following ... anything there look "concerning"?
> >
> > Is this a 2 CPU system? If so, both CPUs are actually running something,
so
> > it is not a deadlock per se.
> >
> >> 99402 www 1 96 0 163M 29892K CPU1 1 0:03 0.00% httpd
> >> 13635 88 34 96 0 92340K 25604K CPU0 0 0:00 0.05% mysqld
>
> Here is what vmstat shows ~10 minutes before (or as) it hung solid last
> time. I didn't think to save the one that ran just before this one (the
> script runs every 5 minutes), but for the 'r b w' columns 'b' was around
> 10ish, while 'w' was 0 ... within a 5 minute period of time, 'w'
> literally skyrockets:
>
> procs memory page disks faults
> cpu
> r b w avm fre flt re pi po fr sr da0 pa0 in sy cs us
sy id
> 107 266 122 16155620 23084 3255 22 1 2 3358 1605 0 0 377 17835
5231 19 7 73
> 6 285 382 16446348 22532 111705 21155 1391 10049 51966 2187328 143 0
36344 499098 423971 3 2 95
> 0 73 386 16440468 23072 7052 1155 85 44 1292 73 372 0 1030 18631
8334 18 12 70
> 0 77 388 16440468 23088 126 1050 0 6 21 27 169 0 521 4186
4125 2 3 94
> 0 66 389 16440468 23104 4 713 0 13 44 58 227 0 352 2217
3504 0 5 95
Well, you had a whole lot of page faults and other VM activity, plus 500k
syscalls. The 'w' is a count of swapped processes, so basically your box is
swapping a whole lot it seems. I think your box is just overloaded.
--
John Baldwin
More information about the freebsd-stable
mailing list